Research On Control Parameters Auto-tuning In Servo-driver System

Servo-driver system is widely used in industry field, especially in large dynamic and high accuracy system such as aerospace and military. Because higher performance is required for the control parameters, so, Servo-driver system control parameters should be always in their best state.More often than not, PID control parameters'setting process is complex and demanding, and the actual running state of servo motor is difficult to observe. Furthermore, actual control effect is difficult to evaluate in different parameters. Therefore, researching of the Servo-driver system control parameters auto-tuning technology have very important engineering practice significance and fundamental value.This paper discusses the law that affect dynamic performance of the Servo-driver system, makes use of the study ability of the BP neural network to tuning the PID control parameters, in order to improve the steady state accuracy and dynamic performance of Servo-driver system.This paper firstly has discussed traditional and its improved PID parameters tuning algorithm. Simultaneously, intelligent PID parameters auto-tuning algorithm was researched. Furthermore, on the basis of researching and analysing artificial neural network parameters auto-tuning algorithm, an improved BP neural network PID auto-tuning method is presented. In the end, the performance of the Servo-driver system is inspected by the theory experiment analysis and simulated by Simulink module under the MatLab platform.The simulation results show: BP neural network PID control parameters auto-tuning control algorithm can overcome the uncertainty and nonlinearity of controlled object, and significantly improve the Servo-driver system's dynamic response, tracking accuracy and reduce the control system's oscillation, achieving non steady-state error in stable state while the rapid dynamic response. From the last known, we can conclude that the BP neural network PID auto-tuning control algorithm proposed is more suitable for the time-varying nonlinearity as well as strong interference in the complex Servo-driver system.